ASWF TAC Meeting - June 16, 2021

Video Conference Link

Voting member attendance

Other Attendees

John Mertic, Linux Foundation
JT Nelson, Pasadena Open Source consortium / SoCal Blender group
Matthieu Mazerolle, Foundry
Patrick Hodoul, Autodesk / OCIO
Nick Porcino, Pixar
Carol Payne, Netflix / D&I WG
Sergio Rojas, Arena World
Scott Wilson, Rust WG
Deke Kincaid, Digital Domain
David Aguilar, Disney Animation

Apologies

Kimball Thurston - Chairperson, Weta Digital Limited
Daniel Heckenberg - Animal Logic Pty Ltd / Industry Representative

Agenda

Slack update
- Working with Slack trying to figure out a plan, trying to get non profit rate but won’t work with our 501.c6 instead of 501.c(3)
- Trying to work on a predictable cost model, will present at governing board meeting next Tuesday
Open Source Days Presentations (make sure to respond to Deb)
- Deb should have sent out communication to project leads
- Email invites to attendees just went out
Review / Approval WG: How’s it going?
- Charter hasn’t really been locked in
- Anyone on this call has something to share?
Project Updates ahead of siggraph
- Any projects trying to make VFX Platform 2022?
- Michael / OCIO: planning a 2.1 release for SIGGRAPH, starting to lock down goals / features for this release. Notified VFX Platform of this version.
- John: ASWF wants to showcase new releases, please share with Rachel, blog post for the new release can be useful.
- Cary / OpenEXR: planning 3.1 releases, coming out as a beta in 2 weeks, official release before SIGGRAPH
- Ken / OpenVDB: released 8.1 a few days ago, targeting 9.0 for SIGGRAPH, not 100% sure yet. 8.1 does not have nanoVDB yet, it’s branched off an old version, so be brought up to 8.1, but still not in main branch, that’s targeted for 9.0. nanoVDB main development being done in a private, on invitation branch, please send GitHub id if interested. Runtime compression can shrink volume size up to 8x without visual artifacts.
- Larry: lots of people are building nanoVDB, but this can make it complicated to line up with VFX Platform, so guidance would be useful. Ken: nanoVDB is part of OpenVDB, but it is also somewhat independent. So when bumping up nanoVDB version, not reflected in OpenVDB. But you can use the one currently in the feature branch. John: nanoVDB is in a private repo? Ken: it is in both. There’s a public version in a feature branch, needs a full review going into main branch. But the active day to day development is happening in a private repo for historical reasons. Larry: should we think of these at separate projects? Ken: no, it will be a single project, and it really would be best to wait for OpenVDB 9.0. John: was asking about private repo since the charters specify that development should be done in public repos? Ken: issue is that some internal NVIDIA developers are committing directly to private repos, this will all be over in a couple of months.
- Scott: Rust WG is going to announce its existence to a wider audience.
- John: please share any major releases with the group and with Rachel for announcements
Lifecycle and working group changes
- Working Group Proposed Changes
  - JF: this seems pretty uncontroversial / no outstanding issues, should we just merge / vote?
  - https://github.com/AcademySoftwareFoundation/tac/pull/242/files#diff-a5ff62b265350989694cd270c55b7b5ef445d0296982b6437334f35aa67ac943
  - A section was added around long term working groups (such as D&I): TAC can establish a WG as a long term WG after a year (requires super majority). WG still has its charter renewed annually, but no expectation that the group needs to be disbanded after a year
  - Sean Looper: should we say “every WG should check in every 6 months”? John: seems reasonable
  - Michael Min: we had also discussed the issue of Committees (long term) vs WG. John: that is an approach we can take. Took the “long term WG” wording to keep things simpler.
  - Sean : what would be the functional distinction between a WG and a Committee? Michael: could be a specific TAC person who sponsors a “Committee” or long term WG.
  - Carol: from perspective of D&I group: words have a lot of meaning, “committee” seems overly formal, “Working Group” seems more descriptive of the work being done. Gordon: if the only distinction is the duration, having a single nomenclature has advantages, especially since you don’t know in advance how long something can last. Also, we don’t want to go a year without hearing back. Committees should report “at least” every 6 months, but can be more often. Sean: can provide this guidance in the charter. Can depend on the nature of the WG, and should be described in the charter of the WG. John: can adjust the wording. Any other concerns? Will then reach out and we can vote to get it in. Gordon: should there be a clause that allows the TAC to vote to amend the charter of a WG? John: typically projects and WGs work quite autonomously, so although the TAC can make requests on a WG, shouldn’t requests to amend the charter come from the WG itself? Gordon: maybe have language that says that TAC can vote on charter amendments, without specifying where the amendments come from. Carol: D&I WG is looking at revisiting its charter after a year, looking at formalizing some specific items around leadership structure. John: it’s useful for WGs to have some introspection about their aims and goals. Have rarely seen friction between TAC and WGs in other projects, in practice it tends to be amicable. Gordon: definitely not proposing any adversarial processes! John: it makes sense to allow a WG to change its charter and propose back to the WG. Will be sent as an email instead of a PR (with the full text).
- Archive Stage
  - Adding more structure / clarity around that stage.
  - Was there something left to adjust? Larry? Make sure that “no longer relevant to the industry” is the criterion for moving a project to Archive Stage.
  - JF: gives TAC an additional tool for deciding that a project is no longer relevant. Bill: and we have the option to move back to adopted stage.
  - JF: propose to accept, Bill: second. All ayes, no opposes / abstains, will be merged in.
- Proposal Process
  - Puts structure around proposal process, and the presentation to the TAC. It has been quite informal so far, now formalizes what a project should prepare for.
  - Changed some “the TAC may” to “the TAC will”.
  - Also discusses annual reviews, and process for the TAC to move the project to the next stage.
  - Bill: we have a backlog of projects to review. Could send the review / proposal ahead over email for review to minimize the time spent in the TAC meetings.
  - Bill: good to adopt this? John: should we vote to adopt the PR? Sean Looper: move to adopt, Michael: second. All ayes, no oppose or abstain.
- Let’s skip the other 2 for now
Project Updates
- OCIO: Michael. Planning for 2.1 release for SIGGRAPH. Just finished SIGGRAPH course which took a lot of time, should be beneficial to the community. JF: can the course material be released? Ken: you sign over the copyright to SIGGRAPH, at least until the end of SIGGRAPH. Carol: should be able to share it after the fact. John: would be interesting to turn into a longer term course, could talk to the training people at Linux Foundation. Michael: it’s a 3hr course, have a 2.5hr video, 200 slide deck. Ken: same for OpenVDB. John: we can discuss after SIGGRAPH. Ken: have given course every 2nd year after SIGGRAPH, have put up the slides on the web site. Would be a good idea to have an introductory course for OpenVDB. JF: training section of aswf.io site? John: you can generate certificates for courses, or at least badges. Patrick: OCIO is planning for a 2.0.2 release (current is 2.0.1)
- OpenEXR: Cary. Working towards significant release prior to SIGGRAPH, mostly work done by Kimball, optimizations in Half type, new threaded implementation of core I/O routines. This weekend is target feature freeze, have all the functionality in place. At the point of starting to formulate message of how to present the new functionality, so that’s what we will be spending the next few weeks on. Also putting out a patch release to 2.5 with security fixes requested by one of the distro package managers. Starting to talk about what we will want to discuss during Open Source Days, trying to accumulate questions we will want to have answers to. Started process of setting up OpenEXR Confluence Wiki page to hold meeting notes, building off what was done for OCIO. Not generally a huge fan of Confluence, but it seems like a nice way to maintain this kind of information without the overhead of storing in the repo, starting to embrace this more.
- OpenVDB: Ken. Released 8.1.0 couple of days ago. Minor release, but major in terms of features and bug fixes, over 100 entries in changelog. Highlights are deprecating support for version 5 (policy is to support only the 2 last major versions listed on VFX Platform). Thinking of getting rid of Python Boost bindings, were added early in the project, but are causing build headaches. Looking to use PyBind11, but no one in TSC wants to do it. Trying to “outsource” the work. If anyone has issues with this, please speak up / reach out. Intending to have an API compatible replacement. Have a SIGGRAPH talk and a long course.
- OSL: Larry. Further integration of SIMD batch shading work from Intel, and work on Optix/CUDA backend. Not quite ready for a release, but not part of VFX Platform, so less schedule pressure.
- OTIO: Nick: focussing on solid Python builds, and delivering Python wheels to PyPI. Going slowly, knocking down problems one by one.
- Asset WG / Eric: framework is up on the Wiki for wishlist requests, encourage each project to think about what production assets would make your project better and post it up there.
- D&I: Carol: reviewing deadline is Friday, mentorships start on July 1st, look out for emails. Should have a further update about accepted 20 students at next meeting.
- Rust/Scott: working on onboarding process, and working on bindings for OpenEXR.
- USD: Michael, had subcommittee meeting with USD Camera, discussing rationale of how metadata is dropped into VFX workflows.